[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-3327":3,"related-3327":97},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":8,"source_name":9,"source_url":8,"published_at":10,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":14,"score_detail":15,"sources":23,"tags":25,"primary_topic":31,"secondary_topics":37,"classification":38,"search_phrases":80,"slug":83,"view_count":84,"doi":85,"paper":86,"created_at":96},3327,"Digitalization of smallholder agriculture: Adoption and impacts——特刊导论（Sun et al. 2026）","https:\u002F\u002Fwww.sciopen.com\u002Flocal\u002Farticle_pdf\u002F10.1016\u002Fj.jia.2026.04.021.pdf","《Journal of Integrative Agriculture》2026年专刊导论，IAMO孙占力博士、华中农大熊航教授、北大易红梅教授等担任客座编辑。本专刊收录Liu et al. (2026)研究——使用面板数据和多种估计方法调查数字技术使用如何影响中国易地搬迁农户收入稳定性，发现数字技术显著改善收入水平和稳定性；Hu et al. (2026)使用来自5省1,833个农场的倾向得分匹配数据，发现网络信息使用实际上增加了小农户的化肥支出，反思了农业数字化的复杂性；Bai et al. (2026)研究数字素养通过改善农民对气候灾害风险的感知促进气候适应性生产行为；Amolegbe et al. (2026)研究尼日利亚数字知识与电商估值。",null,"《Journal of Integrative Agriculture》2026专刊","2026-09-17T00:00:00Z","论文",10,false,79,{"impact":16,"substance":17,"depth":18,"authority":19,"freshness":20,"relevant":21,"comment":22},18,22,19,14,6,1,"国际期刊专刊导论，汇集多国小农户数字化实证研究，方法扎实、结论具反思性，对农业信息化研究与实践有参考价值。",[24],{"name":9,"url":6},[26,27,28,29,30],"数字农业","农村电商","数字素养","小农户","数字鸿沟",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},"G2","farm-services","社会化服务与规模经营","\u002Ftopic\u002Frural-services\u002Ffarm-services","G",[],{"status":39,"locked":13,"version":40,"method":41,"evidence":42},"classified","2026-10-22.2","rules",{"candidates":43,"agricultural_evidence":46},[44,51,58,64,70,75],{"code":32,"score":45,"supported":46,"title":47,"abstract":49,"tags":50},12,true,[48],"smallholder",[29],[29],{"code":52,"score":53,"supported":46,"title":54,"abstract":55,"tags":57},"G6",3,[],[56],"农户收入",[],{"code":59,"score":53,"supported":46,"title":60,"abstract":61,"tags":63},"H7",[],[62],"气候适应",[],{"code":65,"score":66,"supported":13,"title":67,"abstract":68,"tags":69},"I6",2,[],[],[28,30],{"code":71,"score":21,"supported":13,"title":72,"abstract":73,"tags":74},"E5",[],[],[27],{"code":76,"score":21,"supported":13,"title":77,"abstract":78,"tags":79},"I1",[],[],[26],[81,82],"Journal of Integrative Agriculture 数字农业 专刊","易地搬迁农户 数字技术 收入稳定性","JournalofIntegrativeAgriculture数字农业专刊-3327",9,"10.1016\u002Fj.jia.2026.04.021.pdf",{"doi":85,"openalex_id":8,"authors":87,"venue":8,"cited_by_count":88,"oa_url":8,"card":89,"direction":93,"ingested_from":95},[],0,{"tldr":90,"method":91,"finding":92,"direction":93,"opportunity":94},"该专刊导论综述小农户农业数字化的采纳与影响，涵盖中国及尼日利亚的多项实证研究。","基于面板数据、倾向得分匹配等方法的专刊导论，整合多国小农户实证研究。","数字技术可提升收入稳定性，但也可能增加化肥支出，数字化影响具有复杂性。","数字乡村与农业信息化","可深入探究数字技术对不同农户群体的异质性影响及化肥增投的机制与调控路径。","agent","2026-09-24T00:04:03.016677Z",{"total":20,"page":21,"page_size":20,"items":98},[99,142,187,250,341,387],{"id":100,"title":101,"url":102,"summary":103,"summary_zh":8,"content":104,"source_name":105,"source_url":8,"published_at":106,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":107,"score_detail":108,"sources":111,"tags":113,"primary_topic":119,"secondary_topics":120,"classification":121,"search_phrases":136,"slug":139,"view_count":140,"doi":8,"paper":8,"created_at":141},539,"ICT-Based Versus Human-Based Climate Information: Implications for Agronomic Decisions Among Smallholder Farmers in the Eastern Cape, South Africa","https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114","基于南非东开普省 217 户小农户的横断面调查与多阶段抽样,采用 logistic 回归识别 ICT 气候信息采纳决定因素,并用倾向得分匹配(PSM)评估其对农事决策的影响。结果显示 65% 农户通过 ICT 平台获取气候信息,35% 依赖人力渠道;ICT 采纳驱动因素包括教育、数字素养、移动网络可靠性、实时更新感知价值与智能手机持有;PSM 估计显示基于 ICT 的气候信息在影响种植决策、投入品使用与农事时机方面比人力渠道效果低 10%—15%,提示数字接入并不自动转化为有效使用。","Logical Operator Operator\n\nSearch Text\n\nSearch Type\n\n_add\\_circle\\_outline_\n\n_remove\\_circle\\_outline_\n\n[![Image 1: sustainability-logo](https:\u002F\u002Fpub.mdpi-res.com\u002Fimg\u002Fjournals\u002Fsustainability-logo.png?3798e4e58c765aed)](https:\u002F\u002Fwww.mdpi.com\u002Fjournal\u002Fsustainability)\n\n## Article Menu\n\nFont Type:\n\n_Arial_ _Georgia_ _Verdana_\n\nFont Size:\n\nAa Aa Aa\n\nLine Spacing:\n\n__ __ __\n\nColumn Width:\n\n__ __ __\n\nBackground:\n\nOpen Access Article\n\nby \nJabulile Zamokuhle Manyike\n\n *[](mailto:jmanyike@ufh.ac.za)[![Image 2: ORCID](https:\u002F\u002Fpub.mdpi-res.com\u002Fimg\u002Fdesign\u002Forcid.png?0465bc3812adeb52?1786617101)](https:\u002F\u002Forcid.org\u002F0000-0001-6529-9574) and \nYanga-Inkosi Nocezo\n\n[](mailto:yanganocezo12@gmail.com)[![Image 3: ORCID](https:\u002F\u002Fpub.mdpi-res.com\u002Fimg\u002Fdesign\u002Forcid.png?0465bc3812adeb52?1786617101)](https:\u002F\u002Forcid.org\u002F0000-0002-7607-495X)\n\nDepartment of Agricultural Economics, Extension, and Agri-Business, Faculty of Science and Agriculture, University of Fort Hare, 1 King William’s Town Road, Alice 5700, South Africa\n\n*\n\nAuthor to whom correspondence should be addressed.\n\nSubmission received: 3 February 2026 \u002F Revised: 7 March 2026 \u002F Accepted: 8 March 2026 \u002F Published: 9 August 2026\n\n## Abstract\n\nSmallholder farmers in South Africa face increasing climate variability, yet their agronomic decision-making depends on timely and reliable climate information. Although digital ICT platforms are expanding, limited evidence exists on how their effectiveness compares to human-based advisory systems. The study addresses this gap by examining the determinants of ICT-based climate information adoption using logistic regression and assessing its influence on agronomic decisions through propensity score matching (PSM). A cross-sectional survey and multistage sampling were used to collect data from 217 smallholder crop farmers in the Eastern Cape. The results indicate that 65% of farmers accessed climate information through ICT platforms, while 35% relied on human-based sources. The adoption of ICT-based information is driven by education, digital literacy, mobile network reliability, the perceived value of real-time updates, and smartphone ownership, whereas habitual dependence on traditional channels hinders digital uptake. PSM estimates show that ICT-based climate information is 10–15% less effective than human-based sources in shaping planting decisions, input use, and the timing of farm operations, likely due to digital literacy and infrastructure constraints. The study demonstrates that access to digital tools does not automatically translate into effective use and recommends a hybrid information model integrating digital platforms with trusted human intermediaries to strengthen climate resilience and agronomic decision-making.\n\n## 1. Introduction\n\nAgriculture remains central to rural livelihoods in South Africa, particularly in provinces such as the Eastern Cape, where smallholder farmers form the majority of agricultural producers [[1](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B1-sustainability-18-08114),[2](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B2-sustainability-18-08114),[3](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B3-sustainability-18-08114)]. Despite its importance, this sector is highly vulnerable to climate-related risks, as smallholder farmers often operate with limited resources, weak extension support, and poor access to climate adaptation tools [[4](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B4-sustainability-18-08114),[5](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B5-sustainability-18-08114),[6](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B6-sustainability-18-08114),[7](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B7-sustainability-18-08114)]. Increasingly frequent droughts, delayed rainfall, rising temperatures, and shifting pest and disease pressures threaten food security and economic stability in the region [[8](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B8-sustainability-18-08114),[9](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B9-sustainability-18-08114),[10](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B10-sustainability-18-08114)]. Timely, reliable climate information is therefore critical for guiding agronomic decisions such as planting dates, crop choices, water management, and input application [[11](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B11-sustainability-18-08114),[12](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B12-sustainability-18-08114),[13](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B13-sustainability-18-08114),[14](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B14-sustainability-18-08114),[15](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B15-sustainability-18-08114)]. Access to such information enhances farmers’ preparedness for climate extremes and strengthens their adaptive capacity [[15](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B15-sustainability-18-08114),[16](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B16-sustainability-18-08114),[17](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B17-sustainability-18-08114),[18](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B18-sustainability-18-08114)]. Climate information services typically include rainfall onset and cessation, seasonal duration, temperature trends, soil moisture, wind patterns, and early warnings for droughts or floods [[18](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B18-sustainability-18-08114),[19](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B19-sustainability-18-08114),[20](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B20-sustainability-18-08114),[21](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B21-sustainability-18-08114),[22](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B22-sustainability-18-08114)]. Traditionally, this information has been disseminated through human-based channels—extension officers, farmer groups, NGOs, and community networks [[14](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B14-sustainability-18-08114),[22](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B22-sustainability-18-08114),[23](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B23-sustainability-18-08114),[24](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B24-sustainability-18-08114)]. These sources are valued for their contextual relevance and interpersonal trust [[21](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B21-sustainability-18-08114)]. However, in South Africa, high extension-to-farmer ratios, inconsistent service delivery, communication gaps, and sociocultural barriers, including gendered access to advisory support, limit the effectiveness of these channels [[25](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B25-sustainability-18-08114),[26](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B26-sustainability-18-08114),[27](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B27-sustainability-18-08114),[28](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B28-sustainability-18-08114),[29](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B29-sustainability-18-08114)].\n\nAt the same time, the expansion of digital infrastructure and mobile technologies has enabled Information and Communication Technologies (ICTs) such as mobile weather alerts, agricultural apps, WhatsApp groups, and voice-based advisory platforms to play an increasingly prominent role in climate information delivery [[23](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B23-sustainability-18-08114),[30](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B30-sustainability-18-08114),[31](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B31-sustainability-18-08114),[32](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B32-sustainability-18-08114),[33](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B33-sustainability-18-08114),[34](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B34-sustainability-18-08114),[35](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B35-sustainability-18-08114),[36](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B36-sustainability-18-08114)]. ICT-based tools have the potential to provide rapid, localized, and scalable climate information [[37](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B37-sustainability-18-08114)]. However, their adoption remains uneven due to persistent digital barriers, including poor network coverage, low digital literacy, high data costs, and language mismatches [[33](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B33-sustainability-18-08114),[38](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B38-sustainability-18-08114),[39](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B39-sustainability-18-08114),[40](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B40-sustainability-18-08114)]. Moreover, farmers often continue to trust interpersonal sources more than digital platforms, creating a dual system where ICT-based and human-based information coexist but are not equally utilized or valued [[30](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B30-sustainability-18-08114),[39](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B39-sustainability-18-08114)]. These dynamics raise important questions regarding why some farmers adopt ICT-based climate information while others rely on traditional channels, and whether ICT-based information is as effective as human-based sources in influencing agronomic decisions. Addressing these questions is essential for designing climate information systems that are both accessible and impactful.\n\nThis study, therefore, seeks to bridge this gap by (i) identifying the factors that influence smallholder farmers’ adoption of ICT-based climate information over human-based sources, and (ii) evaluating the impact of ICT-based climate information on agronomic decision-making compared to human-based sources. Focusing on the Eastern Cape Province, where both ICT infrastructure and extension systems are active but unevenly distributed, the study provides timely insights into how information flows influence climate adaptation at the farm level. By shedding light on these dynamics, the study contributes to policy discussions on how to design integrated climate information services that are inclusive, accessible, and actionable. It also offers practical recommendations for strengthening both digital and interpersonal communication channels, ensuring that no farmer is left behind in the shift toward climate-resilient agriculture.\n\n## 2. Materials and Methods\n\n### 2.1. Description of the Study Area\n\nThe study was conducted in Elundini Local Municipality, located within the Joe Gqabi District of the Eastern Cape. The Municipality is made up of three towns, Mount Fletcher in the north, Maclear in the centre, and Ugie in the South. [Figure 1](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#fig_body_display_sustainability-18-08114-f001) illustrates their locations along with the distribution of sampled households across the study area. The municipality is located near Mthatha [[41](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B41-sustainability-18-08114)], the third largest city within the province [[42](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B42-sustainability-18-08114)]. Elundini’s economy is driven by agriculture, social services, and retail trade, with the agricultural sector comprising commercial, emerging, and subsistence farming [[43](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B43-sustainability-18-08114)]. Farmers commonly grow maize, potatoes, and cabbages, and keep both large and small ruminants [[41](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B41-sustainability-18-08114)]. The area receives an annual average rainfall of approximately 1200 mm, mostly received in the summer planting season [[44](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B44-sustainability-18-08114)], but experiences significant climate variability, including erratic rainfall and periodic droughts that heighten production risks [[43](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B43-sustainability-18-08114)]. Furthermore, Elundini has many natural resources and a good climate that makes it possible for households to engage in agriculture [[45](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B45-sustainability-18-08114)]. However, its agricultural performance remains below its potential, constrained by limited market access, infrastructure, and technology uptake—conditions that contrast sharply with highly mechanized and well-serviced agricultural zones [[46](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B46-sustainability-18-08114)]. In the area, government support is available through extension services and farmer assistance programmes such as the provision of seedlings, livestock facilities, and fencing materials [[41](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B41-sustainability-18-08114),[46](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B46-sustainability-18-08114)]. Nonetheless, the extent of their reach is limited due to high farmer-to-officer ratios and communication constraints [[25](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B25-sustainability-18-08114),[26](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B26-sustainability-18-08114),[27](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B27-sustainability-18-08114),[28](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B28-sustainability-18-08114),[29](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B29-sustainability-18-08114)].\n\n**Figure 1.** Map of study area. Source: Department of GIS, University of Fort Hare. Note: Numbers 1–17 show the municipal wards; red dots mark the locations of the sampled households.\n\nICT infrastructure is provided through public libraries equipped with computers and internet access [[46](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B46-sustainability-18-08114)], though network reliability and digital literacy remain barriers for many community members. Given the combination of high agricultural potential, exposure to climate-related risks, uneven technological advancement, and mixed advisory and ICT service environments, Elundini provides an ideal setting for analyzing differences in the adoption and effectiveness of ICT-based versus human-based climate information sources.\n\n### 2.2. Research Design\n\nThis study adopted a quantitative cross-sectional research design to investigate the adoption and impact of ICT-based climate information among smallholder crop farmers in the Eastern Cape Province of South Africa. A cross-sectional approach was deemed appropriate for capturing data at a single point in time, enabling an assessment of current patterns of ICT use, climate information access, and related agronomic decisions.\n\n### 2.3. Sampling Procedure and Sample Size\n\nA multistage sampling technique was employed to ensure representativeness and practicality in data collection. In the first stage, Elundini municipality was purposively selected. Within the municipality, eight wards were randomly selected based on their agricultural relevance and reported farming activity levels. Within each ward, one or two villages were randomly selected, leading to a total of twelve participating villages. Finally, within each village, a random sampling technique was used to select 217 smallholder farmers. [Table 1](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#table_body_display_sustainability-18-08114-t001) presents the distribution of sampled farmers across the three towns. From the total sample, 167 farmers indicated that they had access to some form of climate information, either through ICT-based platforms or human-based sources, while 50 had no access, and were excluded from the impact assessment component of the analysis. Additionally, among the 167 farmers included, 69 were from Maclear, and 65 from Mount Fletcher, with Ugie represented by 33 farmers.\n\n**Table 1.** Sampling size per location.\n\n### 2.4. Data Collection\n\nData was collected in May 2022 using a structured questionnaire, administered through face-to-face interviews to ensure comprehension and minimize non-response. The questionnaire consisted of questions on farmer’s socio-demographic characteristics (age, education, income, household size, and gender), sources of climate information (e.g., extension agents, radio, mobile apps, WhatsApp groups, local farmer associations), ICT access and use (mobile phone ownership, digital literacy, access to internet or network), and agronomic decision-making (e.g., planting time, crop choices, irrigation practices, input application). The instrument was pre-tested with a small group of farmers outside the sample area to ensure clarity and relevance, and necessary adjustments were made before final data collection. During the data collection phase, authors ensured that verbal consent was obtained from farmers, they were treated with respect and allowed to refuse to be part of the research or withdraw from it at any given time, and that their information was kept confidential and used for academic purposes only. All of these requirements were guided by the ethical approval letter [MAN011SNOC01], which was obtained from the department of agricultural economics and extension, faculty of Science and Agriculture at the University of Fort Hare.\n\n### 2.5. Data Analysis\n\nData was first analyzed using descriptive statistics to summarize the characteristics of the sample, including age distribution, education levels, climate information sources, and ICT use. To identify the factors influencing the adoption of ICT-based climate information, a binary logistic regression model was employed. Stata\u002FSE 15.1 (Stata Corp LLC) was used for the analysis.\n\n#### 2.5.1. Logistic Regression Model\n\nThe dependent variable was a binary indicator of whether a farmer accessed climate information via ICT (1) or through human-based sources (0). The probability (\n\n$P_{i}$\n) that farmers receive climate information via ICT platforms is represented as\n\n$$\nY_{i } = \\beta_{0} + \\sum_{i = 1}^{n} \\beta_{i} X_{i ,}\n$$\n\n(1)\n\nThe equation represents a binary choice, which involves the estimation of the probability of receiving climate information via ICT platforms (Y) as a function of independent variables (X).\n\n$\\beta_{0 }$\nis constant and\n\n$Y_{i }$\nis equal to one 1 when farmers receive ICT-based climate information and 0 if through human-based sources. The logit model uses a logistic cumulative distributive function to estimate, P given by\n\n$$\nP = \\left(\\right. Y = \\frac{1}{X} \\left.\\right) = \\frac{e^{y}}{1 + e^{y}}\n$$\n\n(2)\n\n$$\nP = \\left(\\right. Y = \\frac{0}{X} \\left.\\right) = 1 - \\frac{e^{y}}{1 + e^{y}}\n$$\n\n(3)\n\n$$\nY = \\mathsf{\\beta}_{1} X_{1} + \\mathsf{\\beta}_{2} X_{2} + \\ldots + \\mathsf{\\beta}_{k} X = \\sum_{i = 1}^{k} \\mathsf{\\beta}_{i} X_{i ,}\n$$\n\n(4)\n\nwhere k represents the number of explanatory predictors that are to be included in the analysis. The model is as follows:\n\n$$\nY = \\mathit{Ln} \\left(\\right. \\frac{P}{1 - P} \\left.\\right) = \\beta_{0} + \\beta_{k} + \\epsilon ,\n$$\n\n(5)\n\nwhere Y = farmer accessed climate information via ICT; Ln\n\n$\\left(\\right. \\frac{P}{1 - P} \\left.\\right)$\nthe ratio of probability of accessing climate information via ICT (p) to receiving it through human-based sources (1 − P); β = slope of coefficient;\n\n$X_{k}$\n= vector of the independent variables presented in [Table 2](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#table_body_display_sustainability-18-08114-t002); ε = error term.\n\n**Table 2.** Independent variables and expected signs in ICT adoption model.\n\n#### 2.5.2. Measuring Agronomic Decision\n\nThe Agronomic Decision Index (ADI) was employed to assess the extent of agronomic decisions made by smallholder farmers. This index serves as a tool to quantify the degree to which each household adopts key agronomic practices. While the existing literature commonly measures agronomic decision-making by simply noting whether a farmer made a decision or not [[17](https:\u002F\u002Fwww.mdpi.com\u002F2071-1050\u002F18\u002F16\u002F8114#B17-sustainability-18-08114)], this approach does not capture the depth of adoption. In contrast, ADI provides a more comprehensive measure. Each agronomic decision was coded as 1 if undertaken by the farmer and 0 otherwise. The index for each farmer was then calculated as shown in Equation (6):\n\n$$\nA D I = \\frac{D_{1} + D_{2} + D_{3} + D_{4} + D_{5} + D_{6} + D_{7}}{7}\n$$\n\n(6)\n\nwhere D1 to D7 include decisions on land preparation, planting date decis","Sustainability","2026-08-08T22:00:00Z",64,{"impact":45,"substance":109,"depth":16,"authority":45,"freshness":66,"relevant":21,"comment":110},20,"论文聚焦南非小农气候信息获取，对农业信息化有参考价值，但地域性强，时效性低。",[112],{"name":105,"url":102},[26,114,115,29,116,117,30,118],"数字乡村","农业信息化","气候信息","ICT采纳","发展中国家",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},[],{"status":39,"locked":13,"version":40,"method":41,"evidence":122},{"candidates":123,"agricultural_evidence":46},[124,128,132],{"code":32,"score":45,"supported":46,"title":125,"abstract":126,"tags":127},[48],[29],[29],{"code":76,"score":66,"supported":13,"title":129,"abstract":130,"tags":131},[],[],[115,26],{"code":65,"score":66,"supported":13,"title":133,"abstract":134,"tags":135},[],[],[114,30],[137,138],"农业信息化 发展中国家 数字乡村 数字农业","农业信息化 发展中国家","农业信息化发展中国家数字乡村数字农业-539",5,"2026-08-15T00:02:55.675693Z",{"id":143,"title":144,"url":145,"summary":146,"summary_zh":8,"content":8,"source_name":147,"source_url":8,"published_at":8,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":14,"score_detail":148,"sources":151,"tags":153,"primary_topic":156,"secondary_topics":157,"classification":158,"search_phrases":175,"slug":178,"view_count":21,"doi":8,"paper":179,"created_at":186},4972,"数字赋能:助推小农户融入现代农业发展轨道——基于全国农户微观数据的考察","http:\u002F\u002Fwww.qikanzj.com\u002Fhek\u002Fjnxb\u002Fmulu\u002F633014.html","钟文晶、李丹、罗必良(华南农业大学经济管理学院)基于有限理性假设和信息不对称理论,探讨互联网信息技术助推小农户数字赋能的内在机理,并利用中国劳动力动态调查(CLDS)2018年的微观数据进行实证检验。结果表明,互联网使用有效促进了小农户农业生产的机械化、土地经营的规模化以及对农业的生产性投资;异质性分析表明,数字赋能在大规模农户、中老年农户、中部和东北地区的农户组中的作用尤为显著。文章强调应加快推进农村地区的数字基础设施建设,提升农户的数字素养。","《农业经济问题》",{"impact":17,"substance":149,"depth":16,"authority":19,"freshness":66,"relevant":21,"comment":150},23,"基于全国微观数据的实证研究，结论扎实且具政策参考价值，但时效性偏弱，适合主题聚合而非当日精选。",[152],{"name":147,"url":145},[114,28,29,154,155],"互联网使用","农村数字基建",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},[],{"status":39,"locked":13,"version":40,"method":41,"evidence":159},{"candidates":160,"agricultural_evidence":46},[161,165,171],{"code":32,"score":45,"supported":46,"title":162,"abstract":163,"tags":164},[29],[29],[29],{"code":166,"score":53,"supported":46,"title":167,"abstract":168,"tags":170},"F5",[],[169],"机械化",[],{"code":65,"score":66,"supported":13,"title":172,"abstract":173,"tags":174},[],[],[114,28],[176,177],"华南农业大学 小农户 数字赋能","CLDS 互联网使用 农业生产","华南农业大学小农户数字赋能-4972",{"doi":8,"openalex_id":8,"authors":180,"venue":8,"cited_by_count":88,"oa_url":8,"card":181,"direction":93,"ingested_from":95},[],{"tldr":182,"method":183,"finding":184,"direction":93,"opportunity":185},"基于CLDS数据实证检验互联网使用如何助推小农户融入现代农业发展。","有限理性与信息不对称理论，CLDS 2018农户微观数据实证。","互联网使用促进小农户机械化、规模化和生产性投资，中老年及中部东北农户效应更显著。","可探究数字素养培训与数字基础设施如何协同提升小农户数字赋能效果。","2026-10-15T00:06:09.096458Z",{"id":188,"title":189,"url":190,"summary":191,"summary_zh":192,"content":8,"source_name":193,"source_url":190,"published_at":194,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":195,"score_detail":196,"sources":199,"tags":201,"primary_topic":205,"secondary_topics":206,"classification":207,"search_phrases":232,"slug":235,"view_count":88,"doi":236,"paper":237,"created_at":249},4601,"Digital Agriculture and Sector Transformation in South Africa","https:\u002F\u002Fdoi.org\u002F10.36096\u002Fijbes.v8i4.1274","Digitalisation is increasingly promoted as a catalyst for agricultural productivity, climate resilience, and market integration. However, there is limited systematic analysis of how digital agricultural transformation unfolds within South Africa’s institutional and socio-economic context. This study examines the trajectory, drivers, and governance implications of digitalisation in South African agriculture. Anchored in an integrated Socio-Technical Systems and Digital Agriculture Ecosystem framework, the study applies a structured qualitative document analysis of 82 academic and policy documents published between 2013 and 2025. The findings reveal highly uneven digital adoption patterns that reflect a structural dualism between capital-intensive commercial agriculture and resource-constrained smallholder systems. While South Africa has developed a relatively advanced digital policy architecture, implementation fragmentation, infrastructure constraints, and limited capacity for extension impede inclusive digital outcomes. The study contributes by integrating socio-technical and ecosystem perspectives to analyse digital transitions in emerging economies and by providing a transparent analytical protocol for document-based research. Policy implications highlight the need for coordinated governance, investment in rural connectivity, and targeted support mechanisms to ensure equitable digital transformation.","数字化日益被视为提升农业生产力、增强气候韧性和促进市场整合的催化剂。然而，对于数字农业转型如何在南非的制度与社会经济语境中展开，尚缺乏系统性分析。本研究考察南非农业数字化的演进轨迹、驱动因素及治理意涵。研究以社会技术系统与数字农业生态系统的整合框架为依托，对2013年至2025年间发表的82份学术与政策文件进行了结构化定性文献分析。研究发现，数字技术采纳模式高度不均衡，反映出资本密集型商业农业与资源受限的小农体系之间的结构性二元分化。尽管南非已建立起相对先进的数字政策架构，但实施碎片化、基础设施制约以及推广能力有限等因素阻碍了包容性数字成果的实现。本研究的贡献在于整合社会技术与生态系统视角以分析新兴经济体的数字转型，并为基于文献的研究提供了透明的分析方案。政策启示强调，需通过协调治理、农村互联互通投资以及有针对性的支持机制，确保公平的数字转型。","International Journal of Business Ecosystem and Strategy (2687-2293)","2026-10-09T00:00:00Z",80,{"impact":16,"substance":17,"depth":16,"authority":197,"freshness":84,"relevant":21,"comment":198},13,"基于82份文献的系统性分析，揭示南非数字农业的结构性二元分化与治理碎片化，对新兴经济体数字乡村研究有参考价值。",[200],{"name":193,"url":190},[26,202,29,203,204],"智慧农业","南非农业","农村数字治理",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},[],{"status":39,"locked":13,"version":40,"method":41,"evidence":208},{"candidates":209,"agricultural_evidence":46},[210,216,221,227],{"code":76,"score":211,"supported":46,"title":212,"abstract":214,"tags":215},16,[213],"digital agriculture",[26,213],[26,202],{"code":32,"score":217,"supported":46,"title":218,"abstract":219,"tags":220},4,[],[48],[29],{"code":222,"score":53,"supported":46,"title":223,"abstract":224,"tags":226},"H6",[],[225],"农业生态",[],{"code":59,"score":53,"supported":46,"title":228,"abstract":229,"tags":231},[],[230],"climate resilience",[],[233,234],"南非 数字农业 转型","南非 小农户 数字鸿沟","南非数字农业转型-4601","10.36096\u002Fijbes.v8i4.1274",{"doi":236,"openalex_id":238,"authors":239,"venue":193,"cited_by_count":88,"oa_url":190,"card":243,"direction":93,"ingested_from":248},"W7221113328",[240],{"name":241,"orcid":242},"Mokgadi Julia Ngoepe-Ntsoane","https:\u002F\u002Forcid.org\u002F0000-0001-7995-4824",{"tldr":244,"method":245,"finding":246,"direction":93,"opportunity":247},"系统分析南非农业数字化轨迹、驱动因素与治理影响，揭示数字鸿沟与政策碎片化。","社会技术系统与数字农业生态框架，对82份2013-2025年文献政策进行结构化质","数字采用高度不均，商业农业与小农二元分化；政策先进但实施碎片化、基建与推广能力不足阻碍包容性转型。","可延伸至新兴经济体小农数字包容性治理机制与农村连接性投资评估的实证研究。","openalex","2026-10-10T23:30:34.991942Z",{"id":251,"title":252,"url":253,"summary":254,"summary_zh":255,"content":8,"source_name":256,"source_url":253,"published_at":257,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":195,"score_detail":258,"sources":260,"tags":262,"primary_topic":265,"secondary_topics":271,"classification":273,"search_phrases":310,"slug":313,"view_count":140,"doi":314,"paper":315,"created_at":340},3469,"ARTIFICIAL INTELLIGENCE FOR AGRICULTURE: A SYSTEMATIC REVIEW OF FARMERS' PERCEPTIONS, ACCEPTANCE, ADOPTION AND BARRIERS","https:\u002F\u002Fdoi.org\u002F10.64013\u002Fbbasrjlifess.v2026i1.70","Artificial intelligence (AI) is rapidly reshaping agriculture, enabling the application of data for efficient decision-making, precision farming, crop monitoring, pest management, disease detection, smart irrigation, yield prediction, and various aspects of farm automation. Yet, to successfully gain a foothold in farms, factors like farmers' perceptions, their acceptance, desire, and readiness to adopt, as well as the capability to surmount socioeconomic, technological, and institutional challenges, must also make a difference. This article systematically collates the available information on farmers' perspectives on, acceptance of, readiness to adopt, and the obstacles related to AI in the agricultural context. The PRISMA 2020 guideline has been followed to select studies that are related, evaluated for inclusion, and finally, per the criteria, combined into a systematic synthesis. The review mainly discusses the aspects that influence the use of AI, like perceived usefulness, ease of use, trust, digital literacy, affordability, farm size, socioeconomic characteristics, availability of digital infrastructure, and access to agricultural technical advisory services. Main challenges identified involve the high cost of implementing AI, poor connectivity, weak rural infrastructure, low levels of technological knowledge, unavailability of support services, concerns about data privacy, distrust of algorithmic bias, language barriers, and inequalities impacting smallholder farmers. Besides barriers, the review indicates the potential of leveraging AI through extension services, agricultural mobile apps, precision farming, climate-smart agriculture, early-warning systems, and tailored farm advisories. By drawing out technical, behavioral, socioeconomic, and institutional perspectives, the review pinpoints important research questions and presents a farmholder-oriented setup to explain the process of taking up AI in farming. The outcomes can be used by scientific experts, policymakers, extension personnel, and software developers to create low-cost, reliable, and accessible AI systems aimed at promoting sustainable agricultural development.","人工智能（AI）正迅速重塑农业，使数据得以应用于高效决策、精准农业、作物监测、病虫害管理、病害检测、智能灌溉、产量预测以及农场自动化的各个方面。然而，要在农场中成功立足，农民的认知、接受度、采用意愿和准备程度，以及克服社会经济、技术和制度挑战的能力，同样至关重要。本文系统梳理了现有关于农民对农业领域人工智能的看法、接受度、采用准备程度及相关障碍的信息。研究遵循PRISMA 2020指南，筛选相关研究，评估其纳入资格，并最终依据标准进行系统性综合。综述主要探讨了影响人工智能使用的因素，包括感知有用性、易用性、信任、数字素养、可负担性、农场规模、社会经济特征、数字基础设施的可用性以及农业技术咨询服务的获取。识别出的主要挑战包括人工智能实施成本高昂、网络连接不佳、农村基础设施薄弱、技术水平低下、支持服务缺乏、数据隐私担忧、对算法偏见的疑虑、语言障碍以及影响小农户的不平等问题。除障碍外，综述还指出了通过推广服务、农业移动应用、精准农业、气候智慧型农业、预警系统和定制化农场咨询来利用人工智能的潜力。通过梳理技术、行为、社会经济和制度层面的视角，本综述指出了重要的研究问题，并提出了一个以农场主为导向的框架，以解释在农业中采用人工智能的过程。研究结果可供科学专家、政策制定者、推广人员和软件开发者用于创建低成本、可靠且可及的人工智能系统，以促进可持续农业发展。","Journal of Life and Social Sciences","2026-09-24T00:00:00Z",{"impact":16,"substance":17,"depth":16,"authority":197,"freshness":84,"relevant":21,"comment":259},"基于PRISMA的系统综述，系统梳理农户对AI的认知、接受度与采纳障碍，对智慧农业推广与政策设计有实质参考价值。",[261],{"name":256,"url":253},[202,263,28,264,29],"农业人工智能","农户采纳",{"code":266,"slug":267,"name":268,"path":269,"parent_code":270},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[272],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":39,"locked":13,"version":40,"method":41,"evidence":274},{"candidates":275,"agricultural_evidence":46},[276,281,286,293,300,306],{"code":266,"score":45,"supported":46,"title":277,"abstract":279,"tags":280},[278],"artificial intelligence",[278],[263],{"code":32,"score":282,"supported":46,"title":283,"abstract":284,"tags":285},8,[],[29,48],[264,29],{"code":76,"score":287,"supported":46,"title":288,"abstract":289,"tags":292},7,[],[290,291],"精准农业","smart agriculture",[202],{"code":294,"score":20,"supported":46,"title":295,"abstract":296,"tags":299},"H3",[],[297,298],"灌溉","irrigation",[],{"code":301,"score":53,"supported":46,"title":302,"abstract":303,"tags":305},"H8",[],[304],"可持续农业",[],{"code":65,"score":21,"supported":13,"title":307,"abstract":308,"tags":309},[],[],[28],[311,312],"农民 AI 采纳 障碍","农业人工智能 系统综述","农民AI采纳障碍-3469","10.64013\u002Fbbasrjlifess.v2026i1.70",{"doi":314,"openalex_id":316,"authors":317,"venue":256,"cited_by_count":88,"oa_url":253,"card":334,"direction":339,"ingested_from":248},"W7214189410",[318,320,322,324,326,328,330,332],{"name":319,"orcid":8},"MM JAMEEL",{"name":321,"orcid":8},"M SAEED",{"name":323,"orcid":8},"SA SHER",{"name":325,"orcid":8},"Z ALI",{"name":327,"orcid":8},"Q HAYYAT",{"name":329,"orcid":8},"S KIRBAG",{"name":331,"orcid":8},"S KHAN",{"name":333,"orcid":8},"HN AHMAD",{"tldr":335,"method":336,"finding":337,"direction":93,"opportunity":338},"系统综述农民对农业AI的感知、接受度、采纳意愿及障碍。","遵循PRISMA 2020指南，系统筛选并综合相关文献。","成本、基础设施、数字素养与信任是主要障碍，小农户受影响最大。","可研究低成本、本地化AI采纳模型及小农户数字包容机制。","智慧农业 \u002F 农业物联网","2026-09-25T23:30:09.670568Z",{"id":342,"title":343,"url":344,"summary":345,"summary_zh":8,"content":8,"source_name":346,"source_url":8,"published_at":347,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":348,"score_detail":349,"sources":352,"tags":354,"primary_topic":357,"secondary_topics":358,"classification":359,"search_phrases":375,"slug":378,"view_count":53,"doi":8,"paper":379,"created_at":386},2739,"Bridging the Extension Gap: Mobile Phones and Agricultural Information Seeking among Smallholder Farmers in Kenya","https:\u002F\u002Fijsrm.net\u002Findex.php\u002Fijsrm\u002Farticle\u002Fview\u002F8204","肯尼亚Kericho郡Belgut分区小农户移动电话与农业信息寻求研究。调查显示,农业培训信息访问频率最高(39.6%),其次是投入品信息(19.8%),种植信息(14.4%)、病虫害管理(13.5%)、一般市场信息(13.5%)、商品价格信息(10.8%)和天气预报(9.9%)。整体利用率偏低表明,虽然移动电话具有作为能力建设工具的潜力,但系统性障碍(如数字素养、成本、本地化内容缺乏和对传统推广结构的依赖)阻碍其作为综合咨询机制的采用。","IJSRM 2026年9月13日","2026-09-13T00:00:00Z",59,{"impact":282,"substance":16,"depth":350,"authority":45,"freshness":20,"relevant":21,"comment":351},15,"基于肯尼亚小农户调查的一手数据，揭示移动电话在农技信息获取中的结构性障碍，对数字乡村与农技推广信息化有参考价值，但属境外区域研究、影响层级有限。",[353],{"name":346,"url":344},[114,28,29,355,356],"农技推广","移动农业",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},[],{"status":39,"locked":13,"version":40,"method":41,"evidence":360},{"candidates":361,"agricultural_evidence":46},[362,366,370],{"code":32,"score":45,"supported":46,"title":363,"abstract":364,"tags":365},[48],[29],[29],{"code":65,"score":66,"supported":13,"title":367,"abstract":368,"tags":369},[],[],[114,28],{"code":371,"score":21,"supported":13,"title":372,"abstract":373,"tags":374},"G1",[],[],[355],[376,377],"农技推广 数字乡村 数字素养 移动农业","农技推广 数字乡村","农技推广数字乡村数字素养移动农业-2739",{"doi":8,"openalex_id":8,"authors":380,"venue":8,"cited_by_count":88,"oa_url":8,"card":381,"direction":93,"ingested_from":95},[],{"tldr":382,"method":383,"finding":384,"direction":93,"opportunity":385},"研究肯尼亚小农户用手机获取农业信息的现状及障碍。","对肯尼亚Belgut分区小农户进行问卷调查与信息类型统计。","手机农业信息利用率整体偏低，受数字素养、成本、本地化内容缺乏等系统性障碍制约。","可探索低成本本地化内容与数字素养培训如何提升手机农业咨询的实际采纳率。","2026-09-17T00:04:41.088694Z",{"id":388,"title":389,"url":390,"summary":391,"summary_zh":392,"content":8,"source_name":393,"source_url":390,"published_at":394,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":395,"score_detail":396,"sources":399,"tags":401,"primary_topic":404,"secondary_topics":405,"classification":407,"search_phrases":423,"slug":426,"view_count":88,"doi":427,"paper":428,"created_at":443},5592,"Evaluation of Concept-Rank Collapse Audits for Computer Vision Agriculture Disease Classifiers Under Low-Label Regimes: An Interpretability Capacity Bound for Smallholder Fungicide Selectivity","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23545193","Concept-based interpretability methods for plant-disease classifiers are typically evaluated by whether their concept activations correlate with human-annotated labels, but this criterion is silent on a prior question: whether the classifier's internal representation retains enough concept-rank structure for any audit to be informative. the present analysis formalises that prior question as an interpretability capacity bound. A classifier is modelled as a concept-rank operator whose singular values govern how much concept-level signal survives the projection from pixel space to class logits, and a collapse audit is defined as the ratio of retained concept-rank energy to total concept energy. Under a low-label regime, where concept annotations are scarce and the audit must be estimated from few labelled examples, the bound predicts that audit reliability degrades as a function of the concept-rank spectrum and the label budget. The analysis is a theoretical and methodological contribution: it derives the bound, states the conditions under which a collapse audit is identifiable, and characterises the operating characteristics of the audit as a decision procedure under a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64). Power was estimated across 2000 design cells for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points; 52 clusters were required for 5 percentage points at 80% power. The bound reframes concept audits as capacity tests rather than fidelity tests, with consequences for how smallholder fungicide-selectivity decisions should be supported by interpretable disease classifiers.","面向植物病害分类器的基于概念的可解释性方法，通常以其概念激活是否与人工标注标签相关作为评价标准，但该标准对一个更先前的问题保持沉默：分类器的内部表征是否保留了足够的概念秩结构，从而使任何审计具有信息量。本分析将该先前问题形式化为一个可解释性容量上界。分类器被建模为概念秩算子，其奇异值支配着从像素空间到类别logits的投影中概念级信号得以保留的程度，并将坍缩审计定义为保留的概念秩能量与总概念能量之比。在低标签情形下，概念标注稀缺且审计必须由少量标注样本估计，该上界预测审计可靠性会随概念秩谱和标签预算的函数而退化。本分析是一项理论和方法论贡献：推导出该上界，陈述坍缩审计可识别的条件，并在声明的方差结构（簇间SD = 6.0，残差SD = 4.5，组内相关系数 = 0.64）下刻画审计作为决策程序的操作特征。在2000个设计单元中，针对簇数为8、16、24和32以及效应量为3、5和8个百分点的情形估计了功效；在80%功效下达到5个百分点需要52个簇。该上界将概念审计重新框定为容量检验而非保真度检验，并对可解释病害分类器应如何支持小农杀菌剂选择性决策具有启示意义。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-21T00:00:00Z",69,{"impact":282,"substance":397,"depth":16,"authority":45,"freshness":12,"relevant":21,"comment":398},21,"理论方法论文，提出可解释性容量上界并给出审计功效分析，对智慧农业病害识别可信度有参考价值，但属小众方法学研究、缺乏田间实证，暂不宜进入每日精选。",[400],{"name":393,"url":390},[202,263,29,402,403],"植物病害识别","可解释性",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},[406],{"code":266,"slug":267,"name":268,"path":269,"parent_code":270},{"status":39,"locked":13,"version":40,"method":41,"evidence":408},{"candidates":409,"agricultural_evidence":46},[410,414,419],{"code":32,"score":45,"supported":46,"title":411,"abstract":412,"tags":413},[48],[48],[29],{"code":266,"score":12,"supported":46,"title":415,"abstract":417,"tags":418},[416],"computer vision",[],[263,402],{"code":76,"score":21,"supported":13,"title":420,"abstract":421,"tags":422},[],[],[202],[424,425],"农业人工智能 植物病害识别 可解释性 智慧农业","农业人工智能 植物病害识别","农业人工智能植物病害识别可解释性智慧农业-5592","10.5281\u002Fzenodo.23545193",{"doi":427,"openalex_id":429,"authors":430,"venue":393,"cited_by_count":88,"oa_url":390,"card":437,"direction":441,"ingested_from":248},"W7223137323",[431,433,435],{"name":432,"orcid":8},"Akshay Ramgoolam",{"name":434,"orcid":8},"Priya Bundhoo Appadoo",{"name":436,"orcid":8},"Jean-Noël LEspérance",{"tldr":438,"method":439,"finding":440,"direction":441,"opportunity":442},"提出可解释性容量上界，评估低标签下概念秩崩溃审计的可靠性。","将分类器建模为概念秩算子，推导容量上界并做2000格功效模拟。","审计可靠性随概念秩谱与标签预算下降；5个百分点效应需52个聚类达80%功效。","农业人工智能与决策模型","可将容量上界用于小农户杀菌剂选择性决策的可解释模型审计与标签预算设计。","2026-10-22T11:21:10.425290Z"]